CPU VPs

To scale for anticipated traffic we increased number of CPU Vps from 15 to 25. Along with that we changed the affinity also. We were expecting increase in performance but it actually degraded the performance. Since then we have tried adjusting PC_HASHSIZE/POOL, DS_HASHSIZE/POOL, LRUs, DD_HASHSIZE/POOL but nothing has helped. Are there any other parameters that we should be looking at too?

On Aug 5, 10:10 am, Mohit <mohitanch...@gmail.com> wrote: > To scale for anticipated traffic we increased number of CPU Vps from > 15 to 25. Along with that we changed the affinity also. We were > expecting increase in performance but it actually degraded the > performance. Since then we have tried adjusting PC_HASHSIZE/POOL, > DS_HASHSIZE/POOL, LRUs, DD_HASHSIZE/POOL but nothing has helped. Are > there any other parameters that we should be looking at too? we are on IDS 11

Mohit wrote: > On Aug 5, 10:10 am, Mohit <mohitanch...@gmail.com> wrote: > >> To scale for anticipated traffic we increased number of CPU Vps from >> 15 to 25. Along with that we changed the affinity also. We were >> expecting increase in performance but it actually degraded the >> performance. Since then we have tried adjusting PC_HASHSIZE/POOL, >> DS_HASHSIZE/POOL, LRUs, DD_HASHSIZE/POOL but nothing has helped. Are >> there any other parameters that we should be looking at too? >> > > we are on IDS 11 > And how many physical CPUs do you have? -- Cheers, Obnoxio the Clown http://obotheclown.blogspot.com

Questions: - How many physical cores do you have? - How fast is the processor clock? - Are you using KAIO for all chunks or AIO VPs? - If AIO VPs do you have enough configured (approx: (1.5 * # non-kaio chunks + 4)) - Do you have enough LRUs and CLEANERS configured to support the increased LRU and buffer contention? - Do you have enough buffers to support more concurrent queries that may be pulling in more data? - Are your RA_ parameters set to high and/or too close together? Art On Tue, Aug 5, 2008 at 1:10 PM, Mohit <mohitanchlia@gmail.com> wrote: > To scale for anticipated traffic we increased number of CPU Vps from > 15 to 25. Along with that we changed the affinity also. We were > expecting increase in performance but it actually degraded the > performance. Since then we have tried adjusting PC_HASHSIZE/POOL, > DS_HASHSIZE/POOL, LRUs, DD_HASHSIZE/POOL but nothing has helped. Are > there any other parameters that we should be looking at too? > _______________________________________________ > Informix-list mailing list > Informix-list@iiug.org > http://www.iiug.org/mailman/listinfo/informix-list > -- Art S. Kagel Oninit (www.oninit.com) IIUG Board of Directors (art@iiug.org) Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Oninit, the IIUG, nor any other organization with which I am associated either explicitly or implicitly. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves.

On Aug 5, 10:43 am, "Art Kagel" <art.ka...@gmail.com> wrote: > Questions: > >    - How many physical cores do you have? >    - How fast is the processor clock? >    - Are you using KAIO for all chunks or AIO VPs? >       - If AIO VPs do you have enough configured (approx: (1.5 * # non-kaio >       chunks + 4)) >    - Do you have enough LRUs and CLEANERS configured to support the >    increased LRU and buffer contention? >    - Do you have enough buffers to support more concurrent queries that may >    be pulling in more data? >    - Are your RA_ parameters set to high and/or too close together? > > Art > > On Tue, Aug 5, 2008 at 1:10 PM, Mohit <mohitanch...@gmail.com> wrote: > > To scale for anticipated traffic we increased number of CPU Vps from > > 15 to 25. Along with that we changed the affinity also. We were > > expecting increase in performance but it actually degraded the > > performance. Since then we have tried adjusting PC_HASHSIZE/POOL, > > DS_HASHSIZE/POOL, LRUs, DD_HASHSIZE/POOL but nothing has helped. Are > > there any other parameters that we should be looking at too? > > _______________________________________________ > > Informix-list mailing list > > Informix-l...@iiug.org > >http://www.iiug.org/mailman/listinfo/informix-list > > -- > Art S. Kagel > Oninit (www.oninit.com) > IIUG Board of Directors (a...@iiug.org) > > Disclaimer: Please keep in mind that my own opinions are my own opinions and > do not reflect on my employer, Oninit, the IIUG, nor any other organization > with which I am associated either explicitly or implicitly. Neither do those > opinions reflect those of other individuals affiliated with any entity with > which I am affiliated nor those of the entities themselves. There are around 30 CPUs and I see kio VPs are same as CPU vps. We already tried increasing LRUs. onstat -F shows that cleaners are not that busy. RA_ has not been changed since last successful run. Would RA_ need to be changed with the change in CPU vp. I see lot of mutex waits on Hash PN_ID so we tried increasing PC_HASHSIZE also, but that didn't help. Is there any limitation with Informix that after certain number of CPU vps it starts to slow down.